Interpretation of results from mutation screening of tumor suppressor genes, such as BRCA2, is becoming an increasingly important part of clinical practice. In most cases, this is quite straightforward, but classification of rare missense variants in these genes presents a difficult problem because it is not known whether these subtle changes in the proteins alter function sufficiently to predispose cells to cancer development. As missense mutations account for approximately 35% of all known variants detected by clinical mutation screening in BRCA1 and BRCA2 this has become a significant clinical genetics issue. Here we propose to evaluate and predict whether 200 relatively common BRCA2 missense mutations are disease causing or neutral using a likelihood model that depends on 1) co-segregation of the mutation with disease in affected families, 2) co-occurrence of the mutation with other known deleterious mutations, 3) family history of cancer, 4) evolutionary sequence conservation, and 5) chemical changes in the amino acid due to the mutation. Having classified a number of the 200 missense mutations, the likelihood model data will be considered as the "gold standard" for disease causality and will be used to establish the sensitivity and specificity of a series of BRCA2 functional assays. The validated functional assays will subsequently be applied to the classification of 50 rare missense mutations from the DNA binding domain of BRCA2 that appears to be enriched for missense mutations. Finally, a more formal statistical approach to estimating the probability that a BRCA2 missense mutation is disease causing will be undertaken using a two-component mixture model that utilizes all of the family, functional, and sequence data. The knowledge gained during the study will improve risk assessment and counseling for carriers of the BRCA2 missense mutations and may lead to better evaluation of other missense mutations in this tumor suppressor gene. Furthermore, the data will provide valuable information about how the DNA binding domain of BRCA2 contributes to cancer.